Dominic Boozle isn't just a professor he's a Pioneer in the fascinating world of human robot interaction as a professor of digital Sciences Robotics and human orentation he explores the future where humans and robots work side by side having LED groundbreaking projects in robotics and AI he's become a leading voice on how Automation and Robotics will reshape Industries and Society itself so what sets him apart his passion for ensuring that robots serve Humanity not replace it expect mindbending discussions about ethics the future of work and the possibilities of robotic [Music] collaboration today I'm joined by Dominic Boozle who explored a future where humans and robots work side by side he's particularly passionate about ensuring that it works alongside humanity and doesn't replace us which I think we can all agree is nice and relaxing to hear so thank you so much for joining me thank you so look first things first are the robots coming yes they are definitely coming but that's nothing to be afraid of why not well uh first of all because they're currently still very limited in their capabilities and uh everybody who has already been working with robots or is working with robots knows that because robots are still programmed line by line by line and we're still far out from autonomous driving or the household robot the robot Butler or whatever that science fiction is telling us and so this is why I'm currently absolutely not afraid but uh have a lot of Hope for Rob a robotic future I would say are we far away from autonomous driving though yes unfortunately yes so I I still remember times when some companies said full autonomous driving in April of 2018 then it became 2020 25 then under certain circumstances and now well everybody predicts that yeah maybe in 2030 on American highways under controlled conditions but so in a in a in a tiny Scottish suburb or a country road or in uh a village in Sicily well that's unfortunately still very far out I I I fear and I would be the first customer you know I'm I'm I'm very much looking forward to the technology but simply not there yet do you think there's a gap between what humans expect robotics to be able to do and then their actual capability like I know I've been in meetings before where we'll literally go through sci-fi films and then think about what the what the robots are doing and then think about how our capabilities match up with them and I know there's a lot of data at the moment that while there's incredible announcements happening in robotics um people do feel frustrated with technology in a way having grown up watching sci-fi films uh yes yes definitely and that was exactly the right right reference because it's in in research even called the Hollywood effect because if I now go out to the streets and ask people about their expectation towards Robotics and artificial intelligence they're always going to site SciFi movies sci-fi books and um usually in two two ways either they're afraid of The Terminator and well also the T-shirt Skynet you know and and these Visions are so more dystopian future or they're hoping for Jetson Rosie R2D2 or all these service Rob Wall-E you know robots taking away uh the the daily chores in making our lives better and easier and unfortunately the the expectation is always if you ask them well if they're positive about it two to three years sometimes five years but sorry for disappointing you I I personally and also the scientific Community more believes like 15 to 20 years so you mention it's about 20 years away I wonder if people think it's coming sooner because of the huge advancements we've seen in AI but it's the real physicality that is stopping us isn't it uh yeah actually both but yeah mostly mostly physicality yeah because uh battery density so all the videos that you're currently seeing about for example humanoid robots those robots actually have a battery lifetime of in in worst case 5 to 6 minutes and in good cases 20 to 30 minutes but not like a 12-hour shift in a in a manufacturing facility so so that is one challenge then they're still very expensive very complex to build um we're still lacking behind so for example if you look at the human arm it's it's a Marvel because with with a very limited weight of 5 to 6 kg you can lift up to well I'm terribly untrained but you could lift up to 10 20 kg or even more if you train your body in in robotics we're still fighting with a payment use usage ratio of uh 1 to2 you know so a 25 kgr robot lifting 10 kg is still an issue sometimes you've worked in robotics for a long time yeah desperately long yes how long uh close to 20 years so 20 years ago when you told your friends that you were working in robotics did they think it was a bit weird oh yeah definitely yeah it was like oh what does that mean so and and actually it was the same thing they they were immediately referring also to science fiction so oh yeah you're building R2-D2 now no unfortunately not we're doing industrial robotics what does that mean well you know the large arms that build cars that do welding that do manufacturing and assembly and it's interesting how little things people actually know about about industrial Robotics and because yeah we always have this this this perception of Science Fiction in our minds and this is unfortunately overexaggerating our expectations I suppose industrial robotics just isn't as sexy as like a Sci-Fi best mate that happens to be a robot you grew up watching exactly and and not really uh for everyday use you know like if you have a robot that weighs four five 600 kilog and yes it can lift half a car and weld it together but that is a very fast efficient unfortunately expensive and poten dangerous machine and this is why um yeah the robots that we're currently seeing in car manufacturing for example that's a robot Zoo they live in EN cages so they are fenced off from humans so in order to protect humans and it's not human robot collaboration that you would expect for example from a service robotic scenario or could you imagine that your robot in the kitchen is behind the fence can you tell me more about that well so so I believe that robotics is actually following four phases the four Rob otic revolutions and the first one was exactly uh robotic automation for car manufacturing assembly production and so on so large machines that are very very efficient very accurate very fast very strong and they can repeat the same task million of millions of times like 1 million two million car doors welding or whatever but unfortunately they're not very flexible and they're not consumer grade and they well because they are lacking sensors and they were not built for it they were not never built for human robot or human machine interaction that's why they're living behind safety enclosures and behind the fence and so yeah and I like to call it the robot zoo or a robot prison yeah if you like like to see like that oh very good one yes I love that um because we're locking them in in and away and we're limiting their flexibility and exactly in iot or industry 4.0 set things that is now the vision to say okay we have flexible lot size one production so down to one part individually made at same cost same efficiency with the same robots but this is not what robotics was built for in the first industrial or in the first robotic Revolution why what was it built for well efficiency output throughput um uh it it repetitive accuracy and especially also diminishing and eliminating very so you wanted every car door to be the same and as as as good as the one before and as good as the one after but you were not looking into individualization or or flexible manufacturing because that was going to be my follow-up question you mentioned that they're not very flexible so why are we locking them away what are the risks well that is usually because uh the robot knows exactly what it's doing and where it's moving but the human might not you know and uh well first of all if something goes wrong okay there's always a very limited uh possibility of failure but uh then it's like okay the robot is always moving at the same path and and doing the same action and if I'm the worker directly working with that work cell I might be aware of that but now imagine that somebody a visitor for example comes in the factory or our worst case always were managers because they're not really working with the machines they're only administering them so um if somebody who doesn't know what the machine does walks in there and then the robot makes an unexpected movement well they're strong enough to push you through the wall and uh that would be terrible and very dangerous and so this is why we actually have to protect humans from machines have there been any incidents of humans being endangered by machines that they created um yes unfortunately yes but um the only cases that I'm I'm aware of over the last 15 years have always involved well actually humans not properly using the machine so for example the these safety enclosures have sensors in the doors to Pro to to ensure that the door is closed while you're working and operating with the machine and uh programming the machine for example so you're holding the the teach pendant and that it has a Deadman switch so if you press to hard or you if you drop it then the robot stops there has been a terrible case where actually somebody died in a car manufacturing plant uh but the reason was that person was working alone which is usually not allowed it should always be two workers um that wor uh person was Bridging the the the sensor in the door with a paperclip so disabled the safety sensor and was performing tasks that were not intended to be done by that robot and it led to a catastrophic failure but it was not the robot failing it was really unfortunately the the the human not following protocol and so these are very very unfortunate events but yeah so in those cases the it's not the fault of the machine because you're very passionate about humans and Machines working together in harmony in this utopian future right yes yeah definitely well I I couldn't imagine any service robotics without humans and robots interacting and but it it is a Stony or has been a Stony path so right now a lot of uh people are talking about collaborative robots or cobots so robots that can interact with humans and uh we now actually have the technology to build robots that have sensors either in their skin or in the joints to detect collisions or avoid collisions so we uncaged the robot in the second robotic Revolution so now we have these machines that can be used for for example working at on the same table with a human or doing assembly tasks together or at least share the same workspace but still not very sexy if you ask me because um the robot is still bolted to the table like this microphone arm here and uh I don't want to go to the robot imagine a household robot where you have to go to the robot Butler and say like listen can you fold my sheets or whatever no the third Revolution will be when those robots become mobile so they should be roaming freely and they should start permeating our living realm and yes it surely will not start in the homes because it's fart difficult but they will from the manufacturing domain they will move into uh retail into offices and there will be more and more tasks so that then finally they might end up in our homes and help us with our daily chores although you know Tesla just announced some robots and they were having a good old walk around yeah oh yeah I know about that and by the way there is a lot of excitement about uh humanoid robots and it's it's not just Tesla so there is also a lot of news of car manufacturers all around the world teaming up with with with uh different companies that built these machines yeah but to be honest unfortunately am I still bit I am still a bit skeptical and the reason is well first of all technicals technical reasons so again energy consumption then complexity cost and if you closely watch those videos they either very repetitive tasks that these robots are performing or or they get a lot of help so for example in those manufacturing demos you always see that there are tiny nuts and bolts coming out of the table to align the part that the robot is working on so it has to be exactly in the same position and same situation and we're still lacking that flexibility or in many cases I would wish for 360 video or 360 degree video because what you're not seeing is the nine people remote controlling that that robot because in most cases this is far from autonomous it is still remote controlled it is remotely operated not by one not by two but by many people well that was the criticism wasn't it yeah that somebody was actually controlling it there's this trend in Tech to announce stuff before it's really even Market ready which is quite interesting so I'm glad we've got you to break it down what what actually is happening um Okay so we've talked about human and robot collaboration sounds nice what kind of tasks are best suited for automation from the robot then and when are the humans going to be getting involved in them well usually um everything that is repetitive um is is something that a robot can do best um everything that needs a lot of accuracy a lot of speed um that is always the same motion and exactly the same position so for example gluing is is one of those those tasks if you're applying glue around a windshield for a car and it has to be exactly the right amount in exactly the right position if you then want to position that windshield into the front of the car that is a typical robot task um human tasks are those where you need either creativity experience knowledge or simply well changing tasks so ah okay now something dropped or now this is not right now I have to correct here or I have to be well uh Flex uh flexible elements like this microphone cable that is something that robots can it is nearly impossible for robots to handle that in a in a production setting whereas for a human plugging plugging this Cable in and out of the socket that's a task of seconds and you can even feel your way so that is the domain of humans don't you think that once generative AI has got even better that that's going to be solved quite quickly oh uh I I I wish I had a dime for every time I had to explain that question to investors because uh so it it's quite funny I I used to have an an interim CTO role for a robot startup for two years and the day we got we got the announcements in from open AI so cat GPT generative Ai and all that large language models I think in that week I had to write 23 or 24 statements for investors why this is not th threatening Robotics and the reason is very simple those AIS are on one hand tailored to a particular task in the in the case of GPT it's made for for handling understanding understanding writing or producing text and speech um but there's nothing for spatial awareness for example so uh to understand if I tell the robot get me that coffee cup uh coffee mug and and how where I can find this in three-dimensional space GPT would be lost also the the image generating AIS would be lost so you would have to build a particular generative AI for those tasks then the data is lacking because we don't have any data like we have the whole internet is full of text and pictures but not of Robotics data and I can give you an example one of the uh world's leading robot manufacturers needed 7 years to finally build a solution for predictive maintenance and it's not because the engineers were not good enough no it's because they were lacking the data the robots were not breaking often enough because industrial robotics has very high expectation toward meantime between failure then on the other hand the customers were not giving them the data from the manufacturing plants because they were concerned that they were giving away inside of knowhow about their their production processes and then thirdly when they had the data the data quality was not good enough in order to do the training of those models so they really had to collect seven years of data in order to come up with such a simple feature and we're not talking about robots that can fre freely roam our homes or Plugin cables so no this is why I'm a bit skeptical about the use of those AI Technologies well you make a good point because essentially all these large language models just scraped the internet for um written data really and image data and it's harder for robots but Tesla have said that they're going to be collecting data from humans did you see that and asking for volunteers that they can just like video record to train how um your body moves so how long do you think that's going to take to train these robots to essentially behave physically like humans yeah remote training of robots is the is is is a little bit the the last result that that everybody is currently re referring to in in that respect because all recording human motion and replicating it and hoping that you could learn from it yeah but uh well Elon Musk gave the answer himself he said those technologies will be available in the long run and in the long run I think he even put a number to it he said at least uh at earliest at 2035 or something like that so it will be another 10 to 15 years he might be a bit optimistic I'm a little bit more pessimistic about but nevertheless so yeah uh recording human motion is a good start but who says that a robotics arm has to move like a human arm or um we now thinking about uh replicating our body with two arms two legs one torso but how about if I have a robot that sits on Wheels and has actually forearms or has an arm with different joints and how can I replicate the data that is taken from a human training the robot and replicate that into a completely different system so this is usually where the problems start but yeah uh first of all telepresence uh teleoperation tele manipulation and recording and using that data is the right direction but unfortunately it's also not the holy grail and it will not solve all those problems do you think consumers and Society are going to be less welcoming to a robot on wheels that has five arms than they are to a robot that looks like a human oh that's a very difficult question because uh it really depends on the on on on what you're using it for and what your expectation is there are studies that claim if you or that prove if you put a face on an industrial robot and make the industrial robot although it doesn't look anything like a human if you make it use make it move in a very human or animall likee way or a very bionic way uh that this immediately uh generates Ates the expectation that this has a soul that this has a behavior that has feeling but it does not it's actually not more intelligent than a screwdriver so um there might be domains where I want the robot to look like a tool because it is a tool and should be treated like a tool and there might be other applications where I want the robot to well cause or or or or Stoke more affection or emotion for example elderly care or something like that and you can work a lot with that so you can make a robot immediately feel very aggressive or very dangerous or very threatening by harsh movements so uh Disney animatronics for example they did did stories uh and and and and studies about that you can take a very large industrial robot and when you train the robot you can make it timid it's like sniffing and feeling and if it touches something something it's pulling back or it's hiding and it immediately becomes something that you want to care for that you want to help that you want to actually enable to learn more but as soon as it is done and it knows what it's doing it can do very large movements be very aggressive very threatening and even with movement you can immediately Stoke those feelings and even more if you give it two legs two arms a face a body and uh well ultimately you might even fall into The Uncanny Valley yeah I mean in some regards in terms of AI we're kind of way past uncanny valley now aren't we but in robotics I guess we've still got a long way to come can you explain what uncanny valley is oh yeah sure uh actually not a place but named after Mr uncanny the the the who discovered this effect is if a robot is looking more and more animal likee or more humanlike our level of acceptance of uh I identification and of personal relation and emotional relation to the system increases it increases up to the point where it gets to 90 95% likeness to the real thing so to an animal to a human and then it drops practically to zero and the reason is because it triggers something that has been embedded in our brains and in our in our in in in our experience for thousands of years which is if a body is moving in a strange way or if a body shows signs of death of sickness so for example zombie or a ghost or something like that then it's danger and then we're immediately afraid and we're put back and then it's either fight or flight so either I try to fight the thing or I try to run away and exactly that response is The Uncanny Valley it's it's also triggered if a robot is not not really human enough but humanoid enough to trigger that effect but not to really jump over the Gap and jumping over the Gap would be if I can't recognize it anymore if it really is a humanoid figurine that I say oh yeah well I don't know if that is now Peter or Robbie mhm it's essentially a sort of fancy Tech word for like a chill down your spine that freaks you out a bit when you think something's human and it's not isn't it yeah exactly but a very strong one that is actually and you can't get over it uh through a n a rational response it's something that is really triggered like uh uh if I if I frighten you and you might start slapping someone that is exactly the same effect so we talked about the shapes of robots and it sounds like the form is really going to follow the function which of course makes sense but I do imagine there's going to be more of a consumer and Society mind shift when it comes to letting in robots that really look too robotic into our homes would you agree well again it really depends um so for example vacuuming robots or floor sweeping robots gutter cleaning robots lawnmowing robots they look nothing like an animal or like a person but we already embrace them and and let them into our homes or into our yards the the only challenge I'm seeing there is they are exactly that they're tools and if we if we continue with that then well soon I think we're going to have 10 15 20 machines roaming around in our house and I'm actually fearing the day that it that it's going to feel like Star Wars you know I I steal a cookie at night I stand in the kitchen eat the cookie something drops five hatchets open seven robots come out and fight who is allowed to clean up and um so so so this is not the way and first of all it's not sustainable and it doesn't make any sense and the more generalistic those robots become and the more tasks they can actually do I think that's when people might want them to look more like well a robot assistant a robot housekeeper a robot helper in any sort and yes then then a humanoid or animal-like shape might help that that task nevertheless uh I think it's not not essentially um it's it's not essentially necessary for people to accept technology yeah it would be bad if the robots not only fight over cleaning up after you but also tell your wife that you were eating CH chocolate biscuits late at night well let's hope not or or depends on which side the robot is then oh you know then we start influencing the robot if more mommies or daddies oh yeah yeah I can imagine um okay so I want to talk about human centered design can you explain what the concept is uh well the the idea of human sensor design is to put the human and and and his or her needs in the in the in the middle of what you're doing so in in focus of what you're doing and I'm absolutely a fan because very often in engineering you see technology push and well I've been working for robotics companies that did exactly that the engineers put in the latest and greatest and the most fanciest uh uh designs and Technologies and whatnot because they wanted to prove that they can sometimes want to show off or they were simply excited because they thought that's what the customer needs unfortunately this was not the case so we ended up with very complex very very high-end but also very expensive machines whereas the customer might have needed a machine for a simple task at a fracture a fraction of the price that does maybe live for two years not for 25 years and so human Center design would exactly look at the hum the the the human needs how to address them how to solve them by the way maybe a robot is not even the best solution for that and the different machine or a different solution can actually be a better choice and then build exactly that and try to address that with the right features I suppose it's just a curse of the technology industry in general that people get so excited about the technology without thinking about why we need it and often it feels like technology for technology sake have you seen that happen a lot in the robotics industry oh robotics iot artificial intelligence well I think every every major hype at a certain point becomes a a self-fulfilling prophecy so uh when I was still in in in with kooka we really had those Germans so so Germany has these small medium Enterprises that are 85% of our economy so the hidden Champions the the So-Cal mle stun and the the midun companies are the ones that produce and manufacturer and we had a lot of them approaching us and they were like Hey listen we uh want to automate and we want to digitize so we have to do industry 4.0 and we were asking yeah cool okay let let's do a project together yeah we're we're happy to help you so uh what do you want to achieve oh we don't know what do you mean you don't know but you just ask us to do a project yes uh about industry 4.0 in iot uh okay what do you want to
achieve oh we also don't know uh who sent you well my boss did okay let's ask the boss so we got the boss into another meeting and then they they we asked him and they said well I don't know either but I read that in the the in the hles blood or or or financial times or wherever and they told me I'm only a fancy and and and upto-date Company if I use those Technologies honestly we were shocked and and this was not just one or two companies this was a lot of them that wanted to look into techn techologies and embrace and use Technologies but not for the added value but for out of political reasons out of the motivation that they thought otherwise they wouldn't be perceived as Innovative companies and so I think this is pretty dangerous because that those Technologies all are great especially Robotics and AI yeah I have to say that but no no they're particularly great but only if you really know why you're using them and what you want to use them for and if they generate added value otherwise it's it's the the investment is not justified so how can business owners who are watching avoid making that same exact mistake the first thing that you have to do is really think about um what do you want to achieve so is it because you want to gain efficiency is it because you want to solve a problem that you couldn't solve before is it because you're running into the global labor shortage or you threatened by the global later short labor shortage by the way we're running out of few labor and the United Nations and the oecd are predicting that this is going to happen around 2045 2050 that's when when we won't find enough people on this planet to do all the work that needs to be done it already starts in some high-skilled areas today um H so if you want to become independent of that and and support your the humans that are working in your production or your manufacturing or what the reason is so it can also be to to be Innovative and to start a tri out project that that is perfectly okay but you should be aware of it then really Define what are your goals and your kpis what is the added value that you want to achieve and then start selecting or or trying out the right Technologies and and and bark into well on that journey is it a coincidence that we're going to run out of workers in factories at around the same time that you predict the robots are coming uh well the the question is what is the hen and what is the egg but uh no actually it is a coincidence because the the dream to come up with artificial workers or artificial helpers is thousands of years old so we find the first proof of automaton so figurines that have been built to take over tasks from humans in ancient Greece and ancient Egypt so there were figurines in temples that were produced using soap pills for ritual Washings after you inserted a coin yeah they were stupid like hell they were it was not robotics and nothing like like we think today but nevertheless they were the the the goal was already there no it's uh just some things are coming together so first of all we from a mechanical point of view we we're slowly getting there and we have I'd say 80% of the Technologies available uh AI is finally taking off so we had two AI Winters now I hope we're littering into AI summer and a very bright AI summer hopefully and so if when these Technologies come come together and when we still work a little bit on them then we can address those issues that are emanating anyways and so it's it's more like a lucky coincidence so it's not that we exactly timed it for that point so that some people were sitting in a basement like oh in 32 years we have to no unfortunately not I mean it just feels like um the robot are primed to take over the lack of humans who will be around um for the factories but let's talk about factories because if the future's robotic we're either going to need loads of new parts for them and I'm wondering how many robots will be in the factories where humans are right now like what does that look like H well first of all uh robot density really depends on the industry that you're talking about and also the the the the way that you're producing um uh the flexibility that you're requesting so as I said plugging in cables is currently still unsolved and would be a very large Market maybe we can tap into this maybe in a second but um then there are other things like car manufacturing and highly highly standardized things but I can tell you what is not going to happen so the the the dream of every manufacturer the so-called dark Factory so Factory where it can turn off the lights because it's 100% robots in there and there's only materials coming in at one end and products coming out at the other end that will not happen simply because it's too complex simply because it's not doable and at a certain point you still need humans to look after the machines to service the machines to operate the machines but robot density might rise and we might have less humans working together with more robots luckily this doesn't mean though that the humans are replaced by machines but they're going to be supported by machines in what way because I think a lot of people are worried that robots might be coming for their jobs oh yes and I'm very lucky to be in the UK today because uh so so were they in the first Industrial Revolution you know uh a lot of of uh Vivas in this country were very uh concerned when the automated weaving machine was invented I think 30,000 or so they went on a large strike and there was a lot of discussion and so on uh there is proof that after 10 years only I think 270 of those 30,000 people didn't find a new job so they were still staying in the industry but they were working with those machines and they were looking after those machines so actually yes it is true if somebody tells you that Robotics and Automation and artificial intelligence is not changing the job landscape then they're lying so it's definitely going to change something but it's not replacing humans humans will have other roles and they will have other focus in their job so not monotonous and repetitive tasks but for example looking after those machines because the first thing that we have to automate is actually the so-called 3D so the dull the dirty and the dangerous jobs because there are a lot of jobs out there that are unergonomic for humans that are monotonous that are repetitive that either hurt your body or your brain and make you mentally unhealthy um or that are too dangerous and or too well simply not suited for humans humans had to do them because let's face it for more than over over for for more more than than than 20,000 years humans were the most versatille machines around and the cheapest machines and they were available in abundance but this is going to change now so um German company Volkswagen has has has numbers that in 2035 from 200,000 workers 30,000 will have retired and they won't find them anywhere on the global job market so 170,000 people will have to do the job of 200,000 so if I want to Future proof myself and although you say it might not be as big of a threat as a lot of media would have us think but if I wanted to make sure that a robot is not going to steal my job what skills should I be developing well first of all you you should be willing to prepare for lifelong learning so um the the future is going to be much more skill-based and um let's me let let me answer it like this my grandfather had one job for 47 years I already had more than 10 jobs in my life my kids that who who are now five and seven years old maybe they will have 50 or 60 jobs in their life for for different companies in different domains maybe slightly changing hopping in and out of an industry and so on so lifelong learning will be a very important skill and and also the capability to um look into a technology understand the technology and find my place in that domain which doesn't mean that everybody has to become a programmer or everybody has to become an engineer but at least I should try to be open for Change and open for new technologies should try to understand them should try to work with them and then find my spot the future of work is changing so rapidly you know yourself compared to your grandfather compared to your kids and the number of job you'll have in that time is so different there's no such thing as a job for life anymore how can educational institutions better prepare students for that future oh yeah very important question um so at my University we're having exactly that discussion because we want to to give the our Bachelor and master students exactly the skills that they need when they're coming out and well it has fundamentally changed I think that it truly is a calculated date but is research that says around 7 1740 so like 250 years ago Humanity had has passed the point where all the knowledge of humanity fit fitted into one head so by 1740 if you were studying mathematics physics philosophy whatever you could learn everything that a human could know after that the knowledge of humanity is much bigger than we can understand and grasp in one Lifetime and it's increasing and it's rapidly increasing and the Cycles are getting shorter and faster so we're running into this problem of over specialization you know you're no longer understanding how a car is built even as an engineer you're understanding how a part is built or how a sensor is built or how that particular sensor is built nevertheless we have to enable them that they have the skills and methodologies to work themselves into new topics that learn new things that push that knowledge into their heads so that they can apply it themselves whereas in the past well we simply had to teach them I'm exaggerating a bit but we had to teach them the formulas and we had to teach them listen this is how you calculate for this gearbox no today I have to explain them the basic knowledge so that they can learn what they need on the daily basis and this is what changed so it's more methodological it's more skill-based and it's more well constru constructivistic learning I wonder if you agree with the idea that it's about Collective human intelligence now especially in the age of AI and Robotics well definitely so Collective human knowledge is um as as as we have more knowledge than one person can know and understand we have to put together those pieces by the way this is one of the reasons why we're struggling with some large projects so for example building a new plane building a new rocket building the humanoid robot of the future is so complex that not one single human or one single team can do it so you need this interdisciplinary approach you need Specialists from all the different areas but unfortunately this is causing new problems so for example that they're not speaking the same language that you have to do a lot of norming and adjusting when you have those multidisciplinary teams you have to get to a common denominator sometimes things are invented in different domains so for example robot Vision well this is a field of research in uh computer science in robotics with the engineers and also in in in the sensory domain so four domains at least and four kinds of researchers are looking into the same problem and this sometimes unfortunately also slows us down as a as a human society with the work landscape evolving so rapidly what initiatives are being taken or should be taken to Res skill the existing Workforce for these kind of changes the experience shows that it's it's actually you can nearly res skill everyone if they want to so the biggest issue is actually the um to to to address those people that are not willing to change so but leaving that aside well the first initiatives are um good skill assessment to understand okay what is my current my current skill set or the current skill set of my employees uh after I know what where they're standing to find out what their potential is so in which direction could or should I develop them by the way also do they want to be developed because some people might say Hey listen if I have four different directions more technical more or more engineering more uh software design Maybe or maybe a more maintenance based approach um and you need different skills for that so if you show them the options and then find out what they want to do they're more motivated to actually follow that path um so skill assessment uh then find out what is the potential then Define individual paths and for sure also identify where lie the Necessities so not that we're today uh to today uh training a lot more Weavers whereas we know that we don't need any Weavers anymore or in that case you know we might not need too many welders right now but on the other hand we need people that can operate welding Machinery or that can can set up a welding manufacturing line um so um I think this is this this is where you should start okay so we don't need Weavers we don't need welders what do we need what should we at least less oh at least less but so what do we need what should we be looking to in the next sort of one to three years in terms of Robotics and The Internet of Things robotics internet of things well what we what we definitely need is people capable of operating and maintaining those machines not everybody has to be an engineer but at least you should be trained in well first of all setting up manufacturing lines and and production lines um then robot programming is we we desperately need robot programmers and we will need them for a long time because it is first of all not too easy well it's not too complex but it is something it's it's a it's a particular way of thinking and and and understanding how you want to replicate a process in the real world so that the robot is able to perform that task so uh Engineers robot programmers uh everything in the field of Maintenance servicing uh and then it really depends on the different domains but yeah in general if you wanted to enter one of those fields technical uh technical understanding would be good uh basic knowledge in programming doesn't have to be really not not everybody has to become a programmer but but scripting and and and those things is important and then uh maintenance tasks what roles do you think there'll be in the future of the robotics industry for people who don't come from a traditionally technical or technology background well again I think it depends on the domain um so for example if you yeah sure if you're working on in in the manufacturing plant and you're not operating the robot and you're not programming the robot well then it might the tasks might be limited but um as I see robots moving out of the manufacturing line getting more into retail spaces service robotics restaurants um uh uh office spaces so there will be uh office automation there will be medical automation there will be uh robots in hospitals and Care Facilities so uh in that case there is a huge amount of tasks that has to be done by people that do have a background in that area so for example I have to know as a lab technician how a blood samples are handled even if they are brought from uh the the patients room to the laboratory by a robot and I have to be able to tell the robot how to get there but then when they arrive I have to put them in the machine I have to operate them correctly I have to maintain temperature I have to make sure that the the right sample is is assigned to the right patient so you still need your domain specific knowledge but now it's interjecting and and interacting with that with robotics technology in um AI some of the biggest use cases have been companionship we've seen um big big uptake of friends therapists life coaches even romantic partners and you know if you look at the clickbait stories from robotics even just a few years ago I mean I wrote one for vice about eight years ago about the robot girlfriends do you see that as being a big market and when it comes to non-traditional roles for people in the robotics industry do you think there'll be something in there for like writers to like make the characters that is a great question um well as we all know that 83% of the internet are actually cat videos um so no not not cat videos but you know um no no uh all joking aside yes I do think that this is that at least there is a large Market the question is if we want or should address it there are already three or four companies on the planet that are building building um what I would at least for this podcast called companionship robots also um sexual companionship robots um and uh they're drowning in pre-orders so there is a wish of people um to use technology also for companionship and for for their their personal needs and and desires nevertheless um there are some ethical and moral challenges attached to that and not just the obvious ones but for example there is um in Japan there is also al already a culture of people living in relationship with dolls and completely nonsexual the the reason is imagine you leave your house every morning at 4:00 a.m. you have to commute for 2 hours you arrive at the factory at 6 you work till 10: you come back at 1 and you repeat there is no room in our our life then for human partnership for human interaction and those people put stuffed so so so so like stuffed animals stuffed dolls at their at their breakfast table in order not to be alone now you could argue it would be great if they had at least an artificial companion on the other hand what would that mean would that in the end potentially mean that we're running into a two class Society where only the rich can afford human interaction and then there is a working class that is limited to robotic or automated uh interaction I I I wouldn't hope to think so but this is one thought the other thought is I'm how do we always say I'm happily married uh but we all know some days more some days less happily uh and but on the other hand what I really appreciate about my marriage is that in many cases my wife and I we don't have the same opinion and we do not agree and there is discussion there is fight there is quarrel there is uh discourse happening imagine now you're living with a robot o AI companion and whenever that companion is not agreeing with you you turn it off or when you fed up with a discussion you turn off the robot and put it in the cupboard what does that do with our human frustration tolerance and we have seen a lot of negative effects of too much remote work in the pandemic now imagine imagine that you would limit your or partially limit your social interaction to robot and artificial companions I don't know if I would like to live in a world like that but that's a personal opinion I had a conversation with a friend who was making a tough love bot to give people the difficult um sort of suggestions and love that they need exactly exactly the opposite of that but I suppose on the other side of the question that I just asked you can you envision a world where robotics actually Foster better humano human interaction oh yeah definitely um it would immediately help uh if for example you would delve into telepresence so imagine I could teleport myself to another place on this planet and interact with people through a telepresence robot and by the way we we're currently in a research project we using these I would call them iPad on Wheels you know like a robot where you just have a screen and even that is better than tele telecommunication through a video call because I can I'm I'm literally sitting at the table I can turn my head I can D address people in the coffee break I can drive out of the room and now imagine I could visit my my grandmother in that way every day I want or I could help her with daily chores or I could tonight even virtually go out uh uh collecting sweets at Halloween with my kids and and and things like that yes definitely then uh everything that is in the field of uh Mobility assistance so robots that help me to stay in my known environment for longer so I don't maybe one day we might have to work longer but at life expectancy is also increasing but I don't want to get directly from work to the retirement home no I want to live 20 30 40 active years and Robotics can help us a lot in our on at at at home and also in our social environment to be capable and to partake in these these environments okay so what should we be prepared for because that is quite a long way away right now people who are watching what should we be looking out for the future of manufacturing the future of Automation and in the future of Robotics well um first of all robotics will it it the the the amount of Robotics in manufacturing and production will increase and this will be classical industrial robotics solutions they will look like machines but they will become more and more versatile they will become more and more intelligent and hopefully one day they will be more like a robot Apprentice hey Robbie have a look uh I want to assemble this and this is how it's put together did you understand that and the robot goes and then does that or I explain simply explain to the robot what the robot should be doing and I don't have to program it anymore so this is what's happening in in industrial robotics then systems will become more flexible they will become mobile they will become smaller and lighter maybe one day humanoid yes but not tomorrow more like the day after tomorrow maybe um from there they will jump over into other domains like service robotics um office robotics retail you maybe we will face 247 supermarkets or service uh inquiries or well plumbers plumber emergency service might no longer be a person showing up but maybe a robot even the telepresence robot imagine that one plumber could be at 10 emergency sites at the same time just logging in and out of the machine but not having to drive there to in order to fix your your leaking leaking leaking water problem um yeah and then from there um hopefully one day household autonomous driving um all these technologies that make our our life easier and better and then one day might free up to again 30% of human time and I just hope that then we are not wasting that time again on cat videos on the Internet or on online gaming but maybe for societal political economic or technical progress and so yeah that we can really make the world a better place do you truly believe that robotics are going to benefit the future of humanity because let's be honest it's not looking great right now on Earth in terms of the future of our population yeah unfortunately not but yes I do believe I'm actually a techno Optimist this is why I'm doing this so um I'm working in robotics because I think Robotics and AI can make the world a better place and I I have an example human population will grow to approximately 10 billion people around 2050 if we want to feed 10 billion people the world food program says we have to look into different kinds of food production and unfortunately production really in the sense of the word which means Urban farming indoor farming vertical farming so there will be plants grown not on soil but on substrate under controlled conditions fully automated in 40 or 50 stories high skyscrapers and on the other hand they will produce the same R of food there will no more there will be no more outbreaks of any uh I don't know eack infections or whatever because there was the wrong fertilizer was used and and the salad was contaminated or so no they will be clean they will have the the same amount of nutrients in there and if you open the packet you will be the first living being to touch that fruit that vegetable that salad on the other hand this will be the way how to actually uh nurture the world in the future and otherwise it will not be possible and um it's not even limited to that because um you might know that there are these these 17 uh sustainable development goals that the United Nations want to tackle by 2035 and you can do that analysis for all of those 17 and uh um there's currently a working group at at the UN um and what we're trying to to analyze is Will Ai and Robotics be able to help us to achieve them quicker or will it even inhibit them and unfortunately for some it's good and for some it's bad so uh one of them is uh equal economic growth especially in the third world and I think it's pretty obvious if if large corporations are dominating the development of technology and they're not giving away their their results for free in form of Open Source then the first world will be growing but the the third world will always be lacking behind and so in those cases we have to look a little bit more into the equal use of technology and well now you can do this for all the 17s or basically like the Chinese menu so the five with seven with double cheese and no onions and then you get into a four-dimensional cube and you find for example out that feeding the world in a sustainable way but also addressing climate action at the same time is a highly complex interaction and Technology will play a role in that but it will be not very very straightforward well that was my next question because I asked about the future of our population but really I think what I wanted to say was how is robotics going to help in protecting the future of our planet because if there's no Planet there is no population exactly exactly and that is one of the biggest challenges uh so well another example is um so we always also have to look into the the the footprint that technology leaves and especially Robotics and artificial intelligence and um people are becoming more and more aware of the fact that AI for example is burning a lot of energy and we have these huge data centers so uh a company like Microsoft Amazon or Google is currently administering um 650 to 850 million servers in that data centers those servers have to be replaced every 18 month because they're burned out this is a huge amount of resources not just talking about uh the energy that we need to supply to those Solutions and they are currently uh uh um they're currently serviced by very high paid engineers and one of the reasons why you have to to swap parts on a 18mon scale is because you don't really know when they're burning out and because you can't react fast enough so if you know brought in technology especially robotics to do the task you could operate those servers longer you could save resources you could make the footprint smaller I just I I still don't say that this is the best way to do it but it is a good intermediary step but then we're back to what we tapped into before uh one of the most important tasks is plugging in network cables and the reason why robots are not doing this today is because they're simply not capable yet and so these are this is how how technology can actually help us with with those challenges and and itself well we we also have to look into ways how to build more sustainable Tech so how to build more sustainable robots that use less resources that are more lightweight that are reusable that are repairable and the same applies to Ai and all the Computing behind it yeah sometimes I feel guilty when I'm asking chat GPT a silly question for wasting the energy um but when it comes to robot ICS are the robot makers considering the impact on the environment luckily yes it's starting more and more so uh in the beginning not it was about how to build the fastest and sturdiest machine possible uh whereas the good thing in industrial robotics has always been that they live very long so usually an industrial robot has a lifetime of 25 to 30 years and then they get refurbished and reused in other countries or other other Manufacturing areas so so that has been a very long lasting technology um but manufacturers put more and more awareness into it like weight saving um only using um materials that are sustainable um ideally getting into different sensor sensory Concepts um avoiding dangerous materials uh uh less energy consumption so for example if you drive an industrial robot in a very aggressive manner then you get energy spikes and that is a problem when you try to supply your manufacturing plant now it's you you can save a lot of energy by softer movement curves like for example a softer acceleration and softer deceleration and also energy recuperation so that you save the energy when you break basically like an in an electric car the the energy that you save when breaking the robot and slowing down the robot can be phased into another step of production to accelerate the robot again and yes they they look into those Technologies so putting some kind of breakes on them so you can't go too fast or too slow um can you explain what a smart sensor is please H yeah um well first of all sensory the sensory landscape is changing in general because um in the beginning when we started in robotics with mobile uh systems and this also applied to the AGS that were roaming around on the manufacturing shop floor they had very expensive sensors like liar you know laser laser sensors laser distance measuring sensors to avoid collisions um same sensors were used in autonomous driving for cars and this is also what made those Solutions so expensive now we're more following um the path of sensory Fusion so the idea is that like humans do we have many sensors we have our eyes our ears our nose tactile sensors in our fingers and so on and the combination of feeling my way around this coffee cup uh a coffee mug and seeing the coffee mug and correcting my movement algorithm by visual servoing it's called so what I'm seeing allows me to have actually a very inaccurate arm because the human arm is very inaccurate but by the use of multiple sensors tactile sensors visual sensors I can use an inaccurate soft piece of hardware and make it very accurate by the use of s of of sensors the same thing is happening there we're now replacing expensive lias by multisensory concept so different lowcost cameras from different angles can replace a 3D sensor for example this is what also Tesla is trying in their cars it's not made for every situation and it's not made for every purpose so there there are always tradeoffs but this is one of the biggest things and the other one is the the other big trend is augmenting and enhancing stupid or lower quality senses by smart software and by smart data analytics so basically filling in the gaps in the data with AI and machine learning and this will also help us a lot in order to work with less perfect sensors to get cost down to increase volume when we increase volume cost goes further down and sensors become more broadly available can you give me a real world example of something that normal people do every single day that is going to be impacted by robotic sensory technology I have a I have a I have a have a nonsensical one which is uh so I usually
2025-01-21 01:22